Re: [RFC PATCH 1/3] filemap: convert page_endio to folio_endio

From: Matthew Wilcox <willy@infradead.org>
To: Pankaj Raghav <p.raghav@samsung.com>
Cc: Christoph Hellwig <hch@infradead.org>,
	hubcap@omnibond.com, senozhatsky@chromium.org,
	martin@omnibond.com, minchan@kernel.org, viro@zeniv.linux.org.uk,
	brauner@kernel.org, axboe@kernel.dk, akpm@linux-foundation.org,
	linux-kernel@vger.kernel.org, linux-block@vger.kernel.org,
	linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
	gost.dev@samsung.com, mcgrof@kernel.org,
	devel@lists.orangefs.org
Subject: Re: [RFC PATCH 1/3] filemap: convert page_endio to folio_endio
Date: Fri, 17 Mar 2023 15:31:53 +0000	[thread overview]
Message-ID: <ZBSH6Uq6IIXON/rh@casper.infradead.org> (raw)
In-Reply-To: <d6cde35e-359a-e837-d2e0-f2bd362f2c3e@samsung.com>

On Thu, Mar 16, 2023 at 11:04:54AM +0100, Pankaj Raghav wrote:
> -       /* clean up. */
> -       while ((page = readahead_page(rac))) {
> -               page_endio(page, false, ret);
> -               put_page(page);
> +               while ((folio = readahead_folio(rac))) {
> +                       folio_clear_uptodate(folio);
> +                       folio_set_error(folio);
> +                       folio_unlock(folio);
> +               }
> +               return;
> +       }
> +
> +       while ((folio = readahead_folio(rac))) {
> +               folio_mark_uptodate(folio);
> +               folio_unlock(folio);
>         }

readahead_folio() is a bit too heavy-weight for that, IMO.  I'd do this
as;

	while ((folio = readahead_folio(rac))) {
		if (!ret)
			folio_mark_uptodate(folio);
		folio_unlock(folio);
	}

(there's no need to call folio_set_error(), nor folio_clear_uptodate())